Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Coding Development for Automated Waveguide Alignment and Direct UV Writing
42
Chapter 4
Coding Development for Automated Waveguide
Alignment and Direct UV Writing
This chapter discusses the approach taken in obtaining automated waveguide alignment
and also in controlling the 3-axis stages for direct UV writing of optical channels. These
coding architectures were developed using the LabVIEW software.
4.1: Brief Introductory on Laboratory Virtual Instrument
Engineering Workbench, LabVIEW
LabVIEW is an acronym for Laboratory Virtual Instrument Engineering Workbench
and a graphical programming language from National Instrument Inc. Instead of using
texts as commands as in traditional programming languages such as C++ and Java, it
uses graphical icons to program. As such, because of this, Labview is referred to as G-
language by most users. LabVIEW is a cross Operating System (OS) platform software
which allows the program created in Windows based OS to work functionally in UNIX
or LINUX environments, and vice versa [1]. This software was developed for the
purpose of measurements, data manipulation and analyzing, and result presentation.
Due to such functionality and flexibility, it is widely applied by researchers to create
their very own virtual instrument for their work [2, 3].
Coding Development for Automated Waveguide Alignment and Direct UV Writing
43
4.2: Automated Waveguide Alignment
4.2.1: Simple Vision System for Automated Waveguide
Alignment
A simple vision system composed of Universal Serial Bus (USB) web-cameras and high
precision alignment stage (Suruga Seiki E4010A) has been built for the purpose of
waveguide alignment. This technique is based on a simple analysis on the fiber and
waveguide position via pictures taken by installed web-cams. The analysis procedure is
simply by introducing a Cartesian coordinate system on picture based on its resolution
with upper left corner of picture as the origin of the coordinate system; hence, a picture
with a resolution of 640x480 has 640 units in abscissa and 480 units in ordinate for
instance. With this designation, coordinates of fiber is defined and thus the stage where
fiber sits can be commanded to move through the commands from computer. Both the
web-cams were set perpendicularly to each other to provide top- and side views for the
alignment; the set up is shown by the Fig. 4.1 and Fig 4.2 below.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
44
Fig. 4.1: Schematic on simple vision system for waveguide alignment
Fig. 4.2: Alignment setup for simple vision system
Coding Development for Automated Waveguide Alignment and Direct UV Writing
45
In this work, picture taken by the top web-cam on closely aligned fiber-waveguide was
first to analyze the respective positions according to the definition mentioned. As soon
as the position of fiber and waveguide were known, fiber is then moved closer with the
waveguide with a fix amount of steps. That was used to determine the change of steps
per pixel within the constant steps of movement and the number of pixels separating
fiber and waveguide. When the separation value is identified, number of steps needed to
move the fiber is therefore determined. Following mathematics demonstrates the
situation mentioned:
Initial position of fiber = F1;
Steps of movement = h;
Second position of fiber after movement = F2;
Waveguide position = W;
Numbers of pixels represent the movement of h value:
F2 – F1 = h
Change of steps per pixel, m:
m = h/ (F2 – F1 ) Eqn. 4.1
Number of pixels separating fiber; k;
k = W – F2 Eqn. 4.2
Steps for fiber to touch the input of waveguide; n:
n = k*m Eqn. 4.3
Coding Development for Automated Waveguide Alignment and Direct UV Writing
46
After the fiber has moved with the steps number as prescribed by Eqn. 4.3, it is directed
to move to the vicinity of input channel of the waveguide with the same principle.
Location of the input channel is only estimation and with assumption that the input
channel was printed precisely at the middle of waveguide. Next, web-cam set at the side
of the alignment setup will take picture from side to confirm the fiber is in not situated
at above nor at bottom of the waveguide. Fig. 4.3 shows the front panel of the simple
vision system and the respective coding is shown in Appendix B.
Fig. 4.3: Front panel of developed simple vision system during operation
Coding Development for Automated Waveguide Alignment and Direct UV Writing
47
4.2.2: Active Alignment Setup Configuration
Apart from the vision simple vision system which had developed in this work,
waveguide alignment has been also carried out through a high resolution stepper motor
(Suruga Seiki E4010A) with the control from a computer via General Purpose Interface
Bus (GPIB) interface to the stepper motor controller (D120), a manual alignment stage
and an optical multimeter. Fig. 4.4 illustrates the experimental setup.
Fig. 4.4: Waveguide alignment setup for channel search
Fiber which is used to delivers optical signal was placed on a high precision
stepper motor alignment stage with six degree of freedoms (DOF). In order to locate the
input channel before the alignment with light detection method; an interference based
method is used to perform the task. A 1-by-4 optical power splitter (silica-on-silica)
sitting on waveguide holder was coupled with an optical signal at the wavelength of
650nm and is scanned vertically through the input facet. The scanning was first started
at one of the upper corner of the power splitter and making the way down to the bottom
of waveguide. During the process, interference fringes were produced from the
Coding Development for Automated Waveguide Alignment and Direct UV Writing
48
waveguide’s output and projected to a screen made of graph sheet via a 40X objective
lens sitting on a manual alignment stage.
During the vertical scan process, number and size of fringes change with
response to different fiber position. At certain when fiber travels from over-clad to
under-clad of the waveguide, number of fringes reduces to a minimum and increase
after that; at the same time, size of each fringe becomes larger up to a maximum and
becomes slimmer slowly. When that turning point of both parameters has reached, light
source is scanned through the input facet laterally to look for the embedded physical
channel in the waveguide which will be represented by a dot on screen. In general,
guiding line represented by a single dot on screen will be the first to meet before the
input channel in which will be represented by multiple dots on screen. As soon as the
position of the input channel has been recognized, fine tune procedure on the input’s
channel position is initiated through active alignment algorithm. Formation of
interference fringes from waveguide’s output may be linked to the structural property of
the waveguide.
When light source is coupled into the over-clad of the waveguide, they were
guided inside the layer and refracted at the output and interfered with each other when
they are leaving from output. Fringes become larger in the aspect of size when the light
is move near to core layer. The moment where the optical signal moves into the core
region, they are refracted at largest angle. As they move away from the core region and
heading towards to under-cladding layer, they refracted at smaller angle similar with
what they had face in the over-cladding layer. The underlying physics on these
phenomena is just the relationship between the angle of refraction and index of
refraction which as described by Snell’s Law. Since the index of refraction of over- and
under-cladding is lower than the core, light will be refracted at largest when it travels
from core to air.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
49
4.2.3: Peak Power Detection
A simple peak power scanning algorithm has been developed for the fine tuning
procedure to locate the optimum fiber position for maximum output power. This
alignment procedure is initiated when the physical property of the waveguide is located
which as represented by bright dots on screen. Basically, this is a semi-automated
algorithm; hence, it was used with the soft-joystick of the stepper controller to perform
the task. Fig. 4.5 shows the front panel of this algorithm during operation and
corresponding coding is shown in Appendix B.
Fig. 4.5: Front panel of peak power detection algorithm
First, this algorithm will perform vertical axis scanning in the waveguide’s input
face for some distance to ensure the input channel will be covered by the scanning
process. Power at any channel is then pick up by the optical power detector which set
for free space measurement and transfer to the algorithm for output display and data
storage. On the completion of the vertical scan, optimum location corresponds with
maximum output power is determined; translational stage is moved to the location using
Coding Development for Automated Waveguide Alignment and Direct UV Writing
50
the soft-joystick then. Immediately after the relocation on z-axis position, this algorithm
is started once again where the scanning will be in y-axis direction in order to locate the
optimum position of maximum output in y-axis direction. Again, the translational stage
will be relocated into the optimum position in y-axis. As soon as the fiber is placed in
the optimum positions determined by this algorithm, output power of each channel is
measured. These output readings are compared to the results obtained from manual
alignment with the soft-joystick.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
51
4.2.4: Manual Alignment on Launch Fiber
A manual alignment technique has also been used to refine the input’s location apart
from the active alignment. This manually method was done by measuring one of the
free space output powers while the translational stage is controlled to move step by step
in z-axis via self-developed soft-key pad. On the earth of a local maximum power is
recorded for a particular z-position, same procedure is repeated for y-axis to look for the
location to which correspond to other local maximum power. The whole process is
repeated up to a point where global maximum power is recorded by the optical power
detector. Once this had achieved, power from each outputs is recorded for the relevant
input wavelengths. The front panel of the developed soft-key pad is showed in Fig. 4.6
and the corresponding coding is shown in Appendix B.
Fig. 4.6 Front panel of soft-key pad of stepper motor controller
Coding Development for Automated Waveguide Alignment and Direct UV Writing
52
4.3: Direct UV writing
In our work, a frequency doubled UV laser operating at 244nm has been used to write
the desired optical channel design. The movement of the sample exposed to the laser
was done via stepper motors and controller from Suruga Seiki. The translational stage is
composed of three individual stages. Circuitry designs were first transferred to pieces of
name cards which have been cut into dimensions similar to the in-house fabricated
samples in order to check that the program functionalities. Fig. 4.7 shows the schematic
diagram for the direct UV writing setup used in this work.
Fig. 4.7: Schematic diagram of direct UV laser writing setup
Alignment of the laser was done step by step by first determined the height of
laser beam from optical table level. Secondly, the first reflective mirror is adjusted the
position so that the beam will hit on the center of the mirror preferably. As soon as the
first mirror is aligned into the position, the same procedure was repeated on the rest of
the reflective mirrors in order to direct the beam to the center of focusing lens.
Attributable to the non-Gaussian profile exhibited by the UV laser, an aperture is used
to trim the beam into a single spot beam by blocking the side modes before the beam is
focused on the translational stage. The focus point was determined by adjusting the
height of the stage relatively to the focusing lens until the tiniest luminescence spot, the
luminescence being due to UV absorption, is observed from the card.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
53
The designed circuitries are 1-by-2 power splitter and Mach Zehnder
interferometer, and the underlying structures of these designs are just straight and
slanted channels. Therefore, the key-point of waveguide writing program is about the
manipulation the two fundamental designs in the Labview algorithm.
The aspects requiring attention in the software development of UV-written
waveguides are the translational speed of stages as this the only parameter having direct
influence on the writing fluence in this case. In order to have a uniform refractive index
change throughout the whole optical circuitry, the corresponding UV writing coding
architecture is preferably designed in such a way where any part of the circuitry design
will not be written twice. In case that cannot be avoided in some particular designs such
as intersection of two channels or either arm of power splitter has to be written twice,
the speed of UV writing has to be increased for the first and subsequent writing
respectively.
Besides from the driving speed denoted by DRV_SPEED (in the unit of pulse
per second, pps) as the main speed controller of the stepper motor, there are another two
parameters were used for the same purpose: start up speed which denoted by S_up
SPEED (in the unit of pulse per second, pps) and rate of acceleration and deceleration
(in ms) denoted by RATE. The pulse in the unit of these parameters represents the
distance unit used in stepper motors which will be µm throughout this work. Since the
stepper motors comprise off electric motors, it is therefore important for the motors to
operate at a relative low speed before they reach the driving speed in the rate of time
specified. That is nothing special but just to prevent accelerated aging of the instrument.
The relationship between these three parameter can be represented by a graph which
shown by Fig. 4.8
Coding Development for Automated Waveguide Alignment and Direct UV Writing
54
Fig: 4.8: Relationship between driving speed (DRI_SPEED), start up speed (S_up SPEED) and rate of
acceleration and deceleration (RATE) [4]
Coding Development for Automated Waveguide Alignment and Direct UV Writing
55
4.3.1: 1-by-2 Optical Splitter
Fig. 4.9: Circuit layout of 1-by-2 optical power splitter
Fig. 4.9 shows the circuitry design of 1-by-2 optical power splitter designed in this work
where the arrows indicate flow of the UV writing process. As shown by the arrows, the
UV writing process was started by performing straight channel writing by process 1 and
this channel design is repeated by process 3, 4, and 7 accordingly. Slanted channel
design writing is continued as soon as process 1 completed, and it is repeated by process
5, and 6. Since there was no electronic controlled shutter being used during UV writing
process, upper arm of circuit is written twice at two-times the speed after process 1 and
this setting continued until process 5. The reason to doubled up the writing speed of
upper arm is to standardized the total dosage of UV receive by the sample.
Particulars on the channel length as well as the value of delta-x and delta-y
which yield the slanted design can be changed within the coding and with
correspondence modification on the stepper motor controller’s resolution. In order for
hassle less operating commands environment, this coding architecture has been
designed to accept the commands of controller’s GPIB address, driving speed, start up
speed and rate only; whereby the direction of where the stage should moved has been
Coding Development for Automated Waveguide Alignment and Direct UV Writing
56
pre-fixed within the coding. Fig. 4.10 shows the interface or the front panel of this
coding architecture.
Fig. 4.10: Front panel of 1-by-2 optical power splitter coding architecture
Basically, this coding architecture is started by moving the translational stage
further away (clockwise direction, CW) from motor which controls the stage movement
in x-direction. As the straight pattern command has been fulfilled, the stage will be
moved in 1µm in CCW of y-direction and 8µm in CW of x-direction; and this
synchronous movement is repeated for 100times in order to produce the slanted pattern.
Next, the stage is moved in CW in x-direction only to complete the intention to create
upper of splitter. Once the arm is created, stage is moved along the same path which
creates the upper arm in counter direction; and is followed by movement of 8µm in CW
of x-direction and 1µm in CW of y-direction concurrently for 100times of iterations.
Again, the stage is moved in CW in x-direction to complete the circuit of 1-by-2 optical
power splitter. The complete coding of this coding architecture is shown in Appendix C.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
57
4.3.2: Mach Zehnder Interferometer
Apart from designing a circuitry for optical power splitting, a circuitry for optical signal
phase modulation is designed as well and is shown by Fig. 4.11.
Fig. 4.11: Mach Zehnder Interferometer layout
First channel of Mach Zehnder interferometer design is started by writing a
straight channel design in the CW direction of x-axis, and is continued by continuous
500times parallel process in x-and y-axis: 8µm in CW direction of x-axis and 1µm in
CCW direction of y-axis; where both of these processes are represented by arrow 1 and
2 respectively. Next, the same instruction of straight pattern of process indicated by
arrow 1 is repeated in process 3. As soon as the process 3 ended, stage will be moved in
8µm in CW direction of x-axis and 1µm in CW direction of y-axis at the same time with
500 of iterations. Starting from this moment, the driving speed of the stage is doubled
up from previous in order for process 5 and 6 to produce a straight design with similar
UV exposed dosage with previous. Soon after that, process 7 in which correspond to the
concurrent movement of stage with 1µm in CW direction of y-axis and 8µm in CCW
direction of x-axis is taken up with the original driving speed. For completion of circuit
Coding Development for Automated Waveguide Alignment and Direct UV Writing
58
design, stage is moved in the CCW direction of x-axis and continued by the slanted
channel design by the 500 instructions of 8µm movement in CCW direction of x-axis
and 1µm movement in CCW direction of y-axis. Again, the steps for the stage to move
can be changed with correspondent change with stepper motor controller’s resolution.
Fig. 4.12 shows the front panel of this circuit design, and the complete coding of
circuitry design is shown in Appendix C.
Fig. 4.12: Front panel design of Mach Zehnder Interferometer coding architecture
Coding Development for Automated Waveguide Alignment and Direct UV Writing
59
4.4: Summary
Throughout this chapter, ideas on the coding architectures of direct UV writing and
waveguide alignment methods have been presented. 1-by-2 optical power splitter and
Mach Zehnder Interferometer are the optical circuitries which have been developed in
this work by transferring the designs into pieces of cards in the size of real waveguides,
and they are compose by series combinations of simple straight channel and slanted
channel designs. The concern on designing the circuitries is to ensure the designs are
able to deliver even UV irradiation to the sample during the writing process, and a
compensation step on the driving speed of translational stage may be useful to unify the
UV irradiation if any part of the circuits needed to be written twice by UV laser. As for
the purpose of pigtailing the 1-by-4 splitter optical power splitter, simple vision system,
interference fringes based alignment, peak power detection algorithm and manual
alignment have been used. The major focus on device alignment was optimizing the
input location of the input channel by detecting the free space output power from output
channels.
Coding Development for Automated Waveguide Alignment and Direct UV Writing
60
References
1. J. Travis, and J. Kring, LabVIEW for Everyone, (Prentice Hall, 2006).
2. V. Denisca, and M. Schreiner, “A LabVIEW-controlled portable x-ray
fluorescence spectrometer for the analysis of art objects,” X-Ray Spectrom. 35,
280-286 (2006).
3. J. H. Moore, “Artificial intelligence programming with LabVIEW: genetic
algorithms for instrumentation control and optimization,” Comput. Meth. Prog.
Bio. 47, 73-79 (1995).
4. D220 stepping motor controller operation manual (Suruga Seiki, 2009).